Deep Learning: GANs and Variational Autoencoders (Udemy.com)

Excelent course, the explanations are very clear and made in intuitive way. Lazy Programmer courses are the best in this subject in this platform. The only issue is that part of the code is written using tensorflow 1.0, and googgle colab right now only support 2.0. One improvement would be to provide implementations using tf2.X too.

Very nice course! The instructor always reply my problems, even they cannot directly provide the plain solution, they still provides some suggestions and concepts of de-bugging! I will finish the rest parts (GANs) later! But I want to leave the message here now!! The sampling techniques were introduced in terms of friendly/systematic way. I can easily obtain the required core concepts to dive in the area of VAE. And I feel very good to implement (partially for some sections like GMM) every mentioned algorithm by myself. While student who are considering to take this course should know that: The style of Lazy Programmer is always that the theory parts are mentioned ONLY in the simplest (easy to be digested) way, the rest parts are practicing sections. Therefore, in this course, practicing is extremely critical to gain knowledge. And most of issues pop up during implementations for every student is different, it's critical to find the problems and try to solve them by going through practicing sections. This course is definitely not for who want to learn things without using their brain!

This is my 12th course by this instructor and as always I learned what I set out to learn. In particular, I enjoyed the variational autoencoder section. I will admit that I skipped through the implementation lectures, as I utilized PyTorch instead. I can recommend the DCGAN tutorial in the PyTorch docs as it closely matches the use case used in Section 4 while taking advantage of PyTorch which I find easier to implement than Theano or Tensorflow 1.x.

Liked this course. It does not contain too much information, so it doesn't get boring. It explains of all the tricky parts so that I'm good to go with own solutions. Code examples could've been split into parts, though, for simplicity. Especially in the GAN and VAE sections. But the course is still great!

Lazy programmer has the ability to teach and give us a thorough understanding of the theory behind machine learning, not just plug in this code and voila. Really appreciate the coding from scratch, it really helps me to understand everything happening from the inputs to the outputs. No stone left unturned. I recommend this course for you if you want to know all the details of GAN and VAE.

I like the Python examples. They help to program in an elegant manner. The presentation in the PP-style is a bit boring. I wished more illustrative explanations about the underlying theory. Implementations of VAs and GANs can be found anywhere, but tutorials written by people who have a deep insight into Variational Inference are very seldom.

Thorough explanation of deep generative models. Assignments are challenging (especially DCGAN) yet can be done with hard work. The lectures are informative and derivations are shown step by step. Instructor is not afraid to use high level math and show all the steps which sets it apart from most online courses.

I think this is the most comprehensive course about GANs in Udemy. However, this is not actually for the beginners, and mostly, it is useful for those who are engaged with advanced machine learning, etc. Me, as a Phd student, could improve my understanding about GANs, and VAE, mostly in a very practical way. Thanks to the instructor.

Very illustrative course on Generative algorithms, which are somehow less well-known in the industry. Code is very clean and direct to the point. Also lessons are crystal clear. It is very satisfactory that the teacher is stressing the stuff that matters without losing the focus on details that are not that important in order to understand how these algorithms work.

This course is great! It provides you with all the necessary knowledge to dive into the exciting world of generative models. The balance between theoretical and implementation details is very good. As ususal in his courses the material is pretty condensed so take your time and take notes! As always a good portion of mathematical background is provided. I found this course much better than the first unsupervised DL part.